All you need is numpy and >= python 3.8
$ python aligner.py <proofread_transcript_id>
This will align the proofread transcript at the given id with the firstdraft word timings.
The output of the alignment is in out/aligned_<transcript_id>.json
usage: align word timings with known transcript [-h] proofread_id
positional arguments:
proofread_id The STELLA transcript id for the proofread transcript
optional arguments:
-h, --help show this help message and exit